﻿README

This repository has the simulation code for:
Hong and Reed (2020), Using Monte-Carlo Experiments to Select Meta-Analytic Estimators
submitted to the Research Synthesis Method


We ran the simulation on a server computer (see below) provided by the Mathematics department at the University of Canterbury.
It took about two months for us to complete the simulation.
In the middle of the simulation, the server computer had to reboot due to important updates, so without interruption it may take less than two months.
> sessionInfo()
R version 3.6.1 (2019-07-05)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows Server x64 (build 14393)


To reproduce our results, a user should run 'Main.R'.
To be able to run, a user must (1) have R and necessary packages installed on his/her machine and (2) set working directory from Main.R file.


Our simulation codes heavily rely on Carter et al. (2019)'s simulation codes that are available from Github.
Carter, E. C., Schönbrodt, F. D., Hilgard, J., & Gervais, W. (2017). Correcting for bias in psychology: A comparison of meta-analytic methods. Retrieved from https://osf.io/rf3ys/



I. Source Files
------------
Main.R - main program (loads all necessary packages and source files)

AR_DGP.R - DGP for Alinaghi and Reed (2018)
AR_simDataset.R	 - Creating/storing datasets for AR simulation
AR_Estimation.R	 - Estimating effects using 11 estimators from AR dataset
AR_CombResults.R - Combining AR simulation results

BR_DGP.R - DGP for Bom and Rachinger (2019)
BR_simDataset.R	 - Creating/storing datasets for BR simulation
BR_Estimation.R	 - Estimating effects using 11 estimators from BR dataset
BR_CombResults.R - Combining BR simulation results

Carter_DGP.R - DGP for Carter et al. (2019)
Carter_simDataset.R - Creating/storing datasets for Carter et al. simulation
Carter_Estimation.R - Estimating effects using 11 estimators from Carter et al. dataset
Carter_CombResults.R - Combining Carter et al. simulation results

SDI_DGP.R - DGP for Stanley et al. (2017)
SDI_simDataset.R - Creating/storing datasets for Stanley et al. simulation
SDI_Estimation.R - Estimating effects using 11 estimators from Stanley et al. dataset
SDI_CombResults.R - Combining Stanely et al. simulation results


1-RE.R	 - Estimating the random effect model and trim and fill
2-P-Curve.R - estimating p-curve
3-PET-PEESE.R - estimating PET-PEESE
4-P-Curve_Skewness.R - estimating skewness
5-P-Uniform.R - estimating p-uniform
6-WAAP.R - estimating WAAP
7-Parameter_Selection_Models.R - estimating 3 and 4 parameter selection model
8-AK_Estimators.R - estimating AK1 and AK2 model
9-EK_Estimator.R - estimating EK model

The folder structure to run the programs should be as follows:

II. Main Folders
import - this folder contains 11 estimators
AR - datasets and estimation results for AR simulation are located in this folder
BR - datasets and estimation results for BR simulation are located in this folder
Carter  - datasets and estimation results for Carter et al. simulation are located in this folder
SDI  - datasets and estimation results for Stanely et al. simulation are located in this folder


III. Sub-Folders
Under each AR, BR, Carter and SDI folder, there are three sub-folders:
simData - this folder contains simulated datasets
analysisResults - this folder contains estimation results
dataFiles - this folder contains aggregated estimation result file

